Oumi Product Update: New Models, Playground & Agent Improvements

It’s been another busy few weeks at Oumi. The platform now makes it easier to build, evaluate, and deploy high-performing custom models.

Compare models against frontier systems

Model Playground lets teams compare fine-tuned models side by side with leading frontier models across performance, latency, and cost. The comparison helps teams make better decisions before moving a model into production.

Expanded model support

New models are available for evaluation and synthetic data generation, including GPT-5.5, GPT-5.6 Luna, GPT-5.6 Sol, GPT-5.6 Terra, Claude Opus 4.7, Claude Opus 4.8, and Kimi-K3.

Gemma 4 E4B Instruct is also supported for training and deployments.

A smarter Oumi Agent

The Oumi Agent now provides stronger guidance, better responses, and improved support across model workflows. Teams can use it to move from an idea to a working evaluation or deployment with less manual setup.